1. The Field of the Invention
The present invention relates to a polarization scrambler that makes the degree of polarization (DOP) of optical carrier zero. For long-haul optical amplifier transmission system, much attention is paid to polarization scrambling in which to improve degradation of transmission performance caused by the polarization dependent gain (PDG) in optical amplifiers, a transmitter scrambles the state of polarization (SOP) of optical carrier for depolarization.
Furthermore, the present invention relates to an optical network configured to input a continuous wave (CW) transmitted via an optical fiber from a light source arranged at one node, to an optical intensity modulator at the other node, modulates the CW using a signal supplied by another network, and transmits the modulated signal light to a next node, wherein a polarization scrambling frequency and a modulation signal bit rate are synchronized to each other by using, taking the polarization dependency of an optical intensity modulator into consideration, polarization scrambling in which the SOP of the CW output from the light source at the one node is scrambled at high speed to depolarize the CW.
Moreover, the present invention relates to an optical network configured to input a CW transmitted via an optical fiber from a light source arranged at one node, to an optical intensity modulator at the other node, modulates the CW using an externally supplied signal, and transmits the modulated signal light to a next node, wherein a polarization scrambling frequency and a modulation signal bit rate are not synchronized to each other by using, taking the polarization dependency of an optical intensity modulator into consideration, polarization scrambling in which the SOP of the CW output from the light source at the one node is scrambled at high speed to depolarize the CW.
2. The Relevant Technology
Polarization Scrambler
In an example of a configuration of a conventional long-haul optical amplifier transmission system, an optical transmitter section is composed of a light source that outputs an optical carrier, an optical modulator that modulates the optical carrier using a data signal to output an optical signal, a polarization scrambler that polarization-scrambles the optical signal, an oscillator that provides the polarization scrambler with a signal with a repetition period, and an optical amplifier that amplifies the polarization-scrambled optical signal. The optical signal output by the optical transmission section is transmitted to an optical receiver section via an optical fiber transmission line and the optical amplifier. The optical receiver section is composed of an optical amplifier, an optical band pass filter (BPF) that reduces the effects of the amplified spontaneous emission (ASE) noise provided by the optical amplifier, and a photodetector that converts an optical signal to an electrical signal to output a received signal.
In this case, the optical amplifier in the transmission line has a polarization dependent gain (PDG) of about 0.1 dB owing to polarization hole burning (PHB). In long-haul transmission system, the effects of PDG are accumulated by multiple optical amplifiers connected together, which fluctuates received optical power. In particular, a decrease in received optical power causes Signal to Noise Ratio (SNR) to be significantly degraded.
(Example of Configuration of First Polarization Scrambler)
The polarization scrambler is an optical phase modulator having a Y axis (phase modulation axis) capable of phase modulation and an X axis which is orthogonal to the Y axis and which is not substantially affected by phase modulation. The oscillator applies a repetition period signal to the optical phase modulator. Incident light traveling through the optical phase modulator in the direction of a Z axis is incident so that its polarization axis lies at an angle of 45° with the X and Y axes. A repetition period signal from the oscillator causes phase modulation only on the Y axis. This causes the SOP of the incident light to be rotated (scrambled) to average the effects of PDG of the multiple optical amplifiers connected together. Consequently, the fluctuation of received optical power can be suppressed.
To maximize the effects of polarization scrambling, all SOP must be allowed to occur uniformly on a time average basis, i.e., the degree of polarization (DOP) must be zeroed. Further, the value for the repetition period of a phase modulated signal has only to be smaller than a time constant (up to 0.1 ms) for PHB in the optical amplifier. In particular, it has been pointed out that polarization scrambling can be improved by using a repetition signal frequency that is about double that of a modulation signal bit rate clock or higher (Reference 1: IEEE Photon. Technol. Lett., vol. 6, pp. 1156 to 1158).
Description will be given of conditions for zeroing the DOP. In general, electrical fields (an electrical field Ex in the direction of the X axis and an electrical field Ey in the direction of the Y axis) having their amplitudes and phases modulated and traveling in the direction of the Z axis can be described as:Ex(t)=a1(t)expi[(ωct−kz)−φ1(t)]  (1)Ey(t)=a2(t)expi[ωct−kz)−φ2(t)]  (2)where ωc, k, and t denote the angular momentum frequency, wave number, and time of the electrical fields, respectively. a1(t), a2(t), φ1(t), and φ2(t) denote a modulated amplitude in the direction of the X axis, a modulated amplitude in the direction of the Y axis, a modulated phase in the direction of the X axis, and a modulated phase in the direction of the Y axis, respectively.
The DOP in output from the polarization scrambler can be varied depending on a phase difference applied to the X and Y axes:δ=φ2−φ1  (3)
The DOP depends on the form of a phase modulation function. However, the condition for achieving a zero degree of polarization if a sine wave is used is:δ=0.7655π sin 2πfsct  (4)
(Reference 5: IEEE J. Lightwave Technol., vol. 8, pp. 838 to 845, 1990). In this equation, fsc denotes a polarization scrambling (rotating) frequency. For the polarization scrambler shown in the first configuration example, described above, this condition corresponds to:φ1=0  (5)φ2=0.7655π sin 2πfsct  (6)(Example of Configuration of Second Polarization Scrambler)
The above described polarization scrambler using an optical modulator has a simple configuration. However, since phase modulation is carried out on only one of the polarization axes, it markedly spreads an optical spectrum. Thus, if a polarization scrambler (optical phase modulator) is used for a wavelength division multiplexing (WDM) transmission system, cross talk may occur between adjacent wavelength channels. This hinders an increase in the spectrum density of WDM signals.
To solve this problem, a polarization scrambling method has been proposed which disperses phase modulation to two axes (Reference 2: Japanese Patent Application Laid-open No. 9-326767 (1997)). With this polarization scrambling method, the phase modulation carried out only on the Y axis (phase modulation axis) is also carried out on the X axis (non-phase-modulation axis), which is orthogonal to the Y axis. Accordingly, repetition period signals provided to the X and Y axes have opposite phases. With this method, the amount of phase modulation applied to the Y axis is dispersed to the two orthogonal axes. Thus, it is possible to accomplish polarization scrambling with the spread of an optical spectrum suppressed.
Specifically, two of the above described optical phase modulators are provided, and a polarization scrambler is arranged between these two optical phase modulators to rotate polarization through 90°. Further, a phase adjuster is used so that repetition period signals applied to the optical phase modulators by an oscillator have opposite phases.
With this configuration, linearly polarized light is incident on a first one of the two optical phase modulators at an angle of 45° with the X and Y axes. Further, light is incident on a second optical phase modulator after having its SOP rotated through 90° so that X axis components that have not been subjected to phase modulation at the first optical phase modulator align with the Y axis. Accordingly, when phase modulation is evenly applied to the phase modulation axis of both modulators with opposite phases, Conditional Expression (4) for a zero degree of polarization changes to:φ1=(0.7655π/2)sin 2πfsct  (7)φ2=−(0.7655π/2)sin 2πfsct  (8)(Example of Configuration of Third Polarization Scrambler)
On the other hand, the polarization scrambling method includes the use of optical short pulses (Reference 3: Japanese Patent Application Laid-open No. 9-326758 (1997)) as well as the above described use of phase modulation. This method comprises associating one bit of a data signal with a plurality of optical short pulses and making the SOP of each optical short pulses within one bit to have different SOP to depolarize the data signal (zero degree of polarization). Depolarization by this method will be described below.
Electrical fields (an electrical field Ex in the direction of the X axis and an optical electrical field Ey in the direction of the Y axis) having their amplitudes and phases modulated and traveling in the direction of the Z axis are expressed by Equations (1) and (2).
On the assumption that components in the direction of the Y axis each have its phase delayed by ε compared to a corresponding component in the direction of the X axis, the intensity I(θ;ε) of light is considered which has passed through a polarizer having a transmission axis in a direction extending at an angle θ with the positive X axis. In this case, an electrical field vector in the θ direction is expressed by:E(t;θ;ε)=Ex cos θ+Eyexp(iε)sin θ  (9)The time average of the intensity is given by:I(θ;ε)=<E(t;θ;ε)E*(t;θ;ε)>=Jxx cos2 θ+Jyy sin2 θ+Jxyexp(−iε)cos θ sin θ+Jyxexp(iε)sin θ cos θ  (10)
In this equation, Jxx, Jyy, Jxy, and Jyx correspond to the elements of the coherency matrix below.
                                 J          =                                    (                                                                                          <                                              Ex                        ⁢                                                                                                  ⁢                                                  Ex                          *                                                                    >                                                                                                  <                                              Ex                        ⁢                                                                                                  ⁢                                                  Ey                          *                                                                    >                                                                                                                                  <                                              Ey                        ⁢                                                                                                  ⁢                                                  Ex                          *                                                                    >                                                                                                  <                                              Ey                        ⁢                                                                                                  ⁢                                                  Ey                          *                                                                    >                                                                                  )                        ⁢                                                  ⁢                                                  =                          (                                                                                          <                                              a                        1                        2                                            >                                                                                                  <                                                                        a                          1                                                ⁢                                                  a                          2                                                ⁢                                                  exp                          ⁡                                                      [                                                          i                              ⁡                                                              (                                                                                                      ϕ                                    1                                                                    -                                                                      ϕ                                    2                                                                                                  )                                                                                      ]                                                                                              >                                                                                                                                  <                                                                        a                          1                                                ⁢                                                  a                          2                                                ⁢                                                  exp                          ⁡                                                      [                                                          -                                                              i                                ⁡                                                                  (                                                                                                            ϕ                                      1                                                                        -                                                                          ϕ                                      2                                                                                                        )                                                                                                                      ]                                                                                              >                                                                                                  <                                              a                        2                        2                                            >                                                                                  )                                                            (          11          )                    
The diagonal elements of J are actual numbers, and the sum of the diagonal elements indicates the total intensity of light.TrJ=Jxx+Jyy=<ExEx*>+<EyEy*>  (12)The non-diagonal elements are generally complex numbers and have the following relationship:Jxy=Jyx*  (13)|Jxy|=|Jyx|≦(Jxx)1/2(Jyy)1/2  (14)
In this case, light with a zero degree of polarization refers to such light that the value for Equation (10) does not depend on θ or ε. A necessary and sufficient condition for this light is:Jxy=Jyx=0  (15)Jxx=Jyy  (16)(Reference 4: M. Born and E. Wolf, Principle of Optics, 4th ed, London: Pergamon Press, 1970, chapter 10.8, pp. 809–816 formula (27)).
The orthogonal polarization components of scrambled light now have equal power, and Conditional Expression (16) is thus met. Accordingly, the DOP is zeroed provided that Conditional Expression (15) is established. That is, provided that there is no temporal overlapping between optical pulses each of which is polarized orthogonally to the succeeding pulse, a1(t)×a2(t) is zero at all points of time regardless of the phase of the optical pulses. It is thus possible to zero the DOP of scrambled light. Reference 3, utilizing optical short pulses, is based on this principle.
However, if there is no temporal overlapping between optical pulses each of which is polarized orthogonally to the succeeding pulse, scrambled light has reduced power. On the other hand, to increase the power of the scrambled light, it is necessary to zero the DOP of the scrambled light in order to cause temporal overlapping between optical pulses each of which is polarized orthogonally to the succeeding pulse. Reference 3 and others have not clarified such a method.
Optical Network
In a basic configuration of a conventional optical network, two nodes are connected together via two optical fibers. In this configuration, a CW output by a light source arranged at a first node is transmitted to a second node via a first optical fiber. An optical intensity modulator at the second node uses a data signal to modulate the CW and transmits the modulated signal light back to a receiver at the first node via a second optical fiber.
Such optical networks include an access network in which a user terminal having no light sources is assumed as a second node and in which an optical carrier supplied by a station (first node) is modulated using a data signal supplied by a user's network (for example, LAN (Local Area Network) and is then transmitted to the station (first node) as an upstream signal. In such an optical network, while an optical carrier passes through an optical fiber in a transmission line, its SOP varies temporarily arbitrarily. An optical intensity modulator used by the user needs to have insignificant polarization dependency. Optical intensity modulators meeting this condition include an electro-absorption (EA) modulator and a semiconductor optical amplifier (SOA) modulator. The use of such an optical intensity modulator enables the construction of the above described optical network.
However, the EA modulator may suffer a heavy transmission loss, so that if it is used to construct the above described optical network, the SNR of a received signal is significantly degraded. Furthermore, in recent years, efforts have been made to develop an EA-DFB laser comprising a DFB laser (Distributed-Feedback Laserdiode) and an EA modulator integrated together. Accordingly, it is very rare that the unitary EA modulator is produced as a device. As a result, it is impossible to expect manufacturing costs to be successfully reduced on the basis of mass production. On the other hand, the SOA modulator does not suffer transmission loss owing to its amplification function. However, it is quicker in response than an erbium doped fiber amplifier (EDFA). As a result, it markedly degrades the waveform of a signal of a Gbit/s order.
On the other hand, an LiNbO3 optical intensity modulator has polarization dependency. However, the above described optical network does not allow the use of a polarization-dependent optical intensity modulator for the previously described reason. In this regard, the polarization-dependent optical intensity modulator can be used in the network if a polarization controller is provided in front of the optical intensity modulator. However, when WDM signals are transmitted, different polarization controllers are required for the respective wavelengths. This is disadvantageous in terms of costs. However, advantageously, the LiNbO3 optical intensity modulator suffers a lighter transmission loss and lower costs than the EA modulator and can carry our modulation at higher speed than the SOA modulator.
Thus, for an optical network such as the one described above in which a light source and an optical intensity modulator are separated from each other via an optical fiber, a configuration has been proposed which enables the use of a polarization-dependent optical intensity modulator such as an LiNbO3 optical intensity modulator (Reference 8: Japanese Patent Application Laid-open No. 2000-196523). This improved configuration solves the problem with the polarization dependency of a second node polarization-dependent optical intensity modulator by arranging, behind the light source, a polarization scrambler and an oscillator that drives the polarization scrambler and regularly rotating (scrambling) the SOP of an optical carrier to zero the DOP of the optical carrier. This configuration requires only one polarization scrambler to transmit WDM signals. It is thus possible to reduce costs compared to the previously described configuration comprising as many polarization controllers as the wavelengths.
The previously described examples of configurations of the first to third polarization scramblers are available as polarization scramblers. However, if an electric signal that drives the polarization scrambler and a data signal bit rate clock are used asynchronously, jitters may occur in a received signal eye diagram (this will be described later). To avoid these jitters, it is necessary to match polarization scrambling frequency with a natural-number multiple of the data signal bit rate clock and synchronize their phases to each other.
However, with the above described improved configuration, it is difficult to match polarization scrambling frequency with a natural-number multiple of the data signal bit rate clock and further synchronize their phases to each other. For example, the first node at which the polarization scrambler is arranged is separated via the optical fiber from the second node at which the polarization-dependent optical intensity modulator is arranged. It is thus possible but not practical to synchronize their phases to each other using a coaxial cable or the like. Accordingly, it is necessary to provide a simple method that achieves phase synchronization regardless of this distance.
Further, to use a polarization scrambler to avoid the polarization dependency of an element inserted into the transmission line as described above, a polarization scrambling (rotation) frequency is used which is double that of the data signal bit rate clock (or higher) (Reference 6: Electron. Lett., vol. 30, pp. 806 to 807, 1994). This is probably due to the sampling theorem that a “signal must be sampled at least twice during each period or cycle of its highest frequency component.”
However, if polarization is rotated at a polarization scrambling frequency which is double that of the data signal bit rate clock (or higher), then modulation causes an optical spectrum to be spread as the polarization scrambling frequency increases. Thus, disadvantageously, the waveform of a signal is dispersed in the transmission line and thus degraded markedly. Further, when WDM signals are transmitted, the spread of the optical spectrum may hinder an increase in the spectrum density of WDM signals. The polarization scrambling frequency is desirably set to be lower than the double of that of the data signal bit rate clock.
To realize this, a method has been proposed in which for a polarization scrambler of the first or second configuration, the data signal bit rate clock is synchronized to a sine wave electric signal driving the polarization scrambler, to set the polarization scrambling frequency equal to that of the data signal bit rate clock, thus narrowing the spectrum of output light from the polarization scrambler (Reference 7: Electron. Lett., vol. 32, pp. 52 to 54, 1996). However, with this method, it is essential to synchronize the phase of the data signal bit rate clock to the phase of the sine wave electric signal driving the polarization scrambler. Consequently, a simple phase synchronization method is required as in the case in which jitters must be avoided.
Thus, the two points described below can be given as guidelines for implementing an optical network using the above described improved configuration.
{circle around (1)} To obtain a received signal that is not degraded by jitters, certain measures must be taken to synchronize the phase of the data signal bit rate clock to the phase of the electric signal driving the polarization scrambler. This phase synchronization method also contributes to narrowing the spectrum of output light from the polarization scrambler of the first or second configuration.
{circle around (2)} The received signal is allowed to be degraded by jitters. The electric signal driving the polarization scrambler and the data signal bit rate clock are set to be asynchronous. The polarization scramble frequency is set to be lower than the double of that of the data bit rate clock to narrow the spectrum of output light from the polarization scrambler. This method requires no phase synchronization means and enables the network configuration to be simplified compared to {circle around (1)}.